Idea Creator

wanshuiyin/auto-claude-code-research-in-sleep

Generates and ranks publishable research ideas from a broad direction. It runs a three phase pipeline: surveys recent literature and your local paper library to map the landscape, fans out candidate generation across multiple analytical lenses (method transfer, contradictions, untested assumptions, scaling regimes), then routes everything through a cross-model jury using Codex or manual review backends to filter for novelty and feasibility. Includes optional micro-pilots (GPU budget capped at 8 hours total) to validate top ideas with code before you commit. Integrates with the research wiki if present to avoid regenerating failed ideas and leverage known gaps. The reviewer routing is strict about not letting Claude judge Claude, which adds setup friction but keeps the quality gate honest.

skills.sh installs: 609
GitHub stars: 18.9K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add wanshuiyin/auto-claude-code-research-in-sleep --skill idea-creator --agent claude-code

Installs into .claude/skills of the current project.

Files

Research Idea Creator

Before each reviewer slot, follow Selectable Reviewers: resolve invocation → .aris/reviewer.json → Claude Code default (Codex MCP) with reviewer_route.py through the canonical helper resolver, preserving this slot's tier, fresh contexts, source/image inputs and deterministic gates. Codex examples/pins below apply only to backend=codex; other routes use their own first-call/reply contract. Save the route and actual returned identity; unavailable selections are REVIEW_UNAVAILABLE. Classify acceptance from actual families/completion/scope. Forward invocation reviewer/model/effort overrides to reviewer-bearing children. Generation workers are unchanged.

Generate publishable research ideas for: $ARGUMENTS